Standard Operating Procedure Bruker Quest Diffractometer with copper wavelength X-ray microsource

Transcription

1 Standard Operating Procedure Bruker Quest Diffractometer with copper wavelength X-ray microsource X-ray Crystallography Laboratory Purdue Department of Chemistry The following is a guide for collecting data and solving structures using the Purdue Bruker Quest single crystal diffractometer with a copper wavelength X-ray microsource. It is intended as a walk-through user guide geared especially towards novice users, but also tries to cover more advanced features of the software where they are important for the collection of normal simple small molecule structural data. A basic knowledge of the fundamentals of diffraction and crystallography is expected. This manual is based on an instrument and software produced around the year Most of the manual s content also applies to newer as well as older instrument makes. For the novice user, changes between different generation instruments are mostly limited to slightly different program layouts and color schemes. The general procedures described in this manual still apply. For a more in-depth description of the features of a CCD or CMOS diffractometer, the reader should refer to the manuals and technique guides on specific topics by the manufacturer of the type of instrumentation you are using. The gold standard for a more in depth guide towards the use of Shelxtl for the refinement of single crystal structures is Peter Müller s book Crystal Structure Refinement. Every crystallography lab should have at least one copy. Among the many programs commonly used for crystal structure solution and refinement we recommend the Bruker Shelxtl package (including XPREP, XS and XM), George Sheldrick s refinement program Shelxl2017, the graphical interface Shelxle by Hübschle, Dittrich and Sheldrick, and Platon by Anthony Spek. 1

2 Table of Contents Instrument Overview... 3 Adjusting the Temperature... 4 Starting the System... 5 Unloading the Previous Sample... 6 Selecting and Mounting of Sample... 7 Crystal Mounting and Centering... 9 Crystal and Compound Description Unit Cell Determination Data Collection Data Integration Scaling and Absorption Correction Transferring the Data Setting up Data for XPREP XPREP The Initial INS File Solving Structures using XS and XM Structure Refinement

3 Instrument Overview The Bruker Quest instrument with a copper microsource consists of: The diffractometer including the X-ray enclosure, detector, mircosource IµSCu X-ray tube, multilayer optics for monochromatization and collimator, video microscope, kappa (4-axis) goniometer, electronic controls, power supply and other miscellaneous pieces of equipment essential for the operation of the instrument. An Oxford Cryosystems 800 variable temperature unit (pump, controllers and liquid nitrogen tanks located outside the enclosure) and an Oxford Cryosystems AD51 dry air unit. The instrument computer: This computer runs the Measurement Server, the Bruker Instrument Service (BIS), Diffrac.Maintenance and Apex3. Figure 1, Inside of the Quest X-ray enclosure with microsource X-ray tube, Detector, Variable Temp Unit and Goniometer. Between X-ray tube and detector are the Optics (Multilayer Optics for monochromatization, Collimator), the Video Microscope (in back, partially hidden), and the Beam Stop. On the Goniometer is mounted the Goniometer Head holding the crystal specimen. 3

4 Having a promising sample of crystals, it is best to start in the following order: Adjusting the Temperature All but very high melting samples (inorganics, ceramics, etc) should be measured at lower temperature to avoid extensive thermal motion of the atom cores, to obtain higher angle diffraction data and to minimize radiation damage to the sample. The Oxford Cryosystems variable temperature unit is a stand-alone unit, operated through its own controllers. It can be also controlled via Apex3 or BIS. The sample is embedded in a stream of cold nitrogen gas supplied by a low pressure liquid nitrogen dewar. To avoid buildup of ice, the sample is insulated from ambient humidity by an outer layer of warm dry air supplied by an Oxford Cryosystems AD51 dry air unit (located outside the lab in a facilities closet). The Oxford Cryosystems variable temperature unit is able to achieve temperatures between 375 K ( C) down to ca. 85 K ( C). The recommended temperature for low temperature data collections is 100 K in winter, and 150 K in summer (due to higher humidity levels that can cause icing around the sample mount). Make sure, the tank is properly connected and full enough for the planned experiment. The Oxford Cryostream N2 tank (silver dewar behind the diffractometer) is automatically refilled from the larger low pressure liquid nitrogen tank (between the two diffractometers). Check that the Liquid Level Controller (the smaller of the two controllers, Fig. 2) is set to <AUTO> for automatic refills. The fill level displayed on the controller should be between 20 and 80%. When it is below 20% (despite the controller set to AUTO) the low pressure N2 tank is empty and needs to be refilled. 4

5 Figure 2, The external controller (bottom) and the circulation pump (top) of the Oxford Cryosystems variable temperature unit. Check that no data collection is running (e.g., look at the BIS interface, check for Frames left or Time left ). If a running data collection does not need to continue abort the data collection (see below), then make changes to the VT unit. Check that the dry air unit providing the shield gas flow is switched on. Check on the display of the Cryostream Controller (larger of the two controllers) if the variable temperature unit might be already running. If the unit is running check if the temperature is appropriate for your experiment. If not, use the <Run Experiment> interface of Apex3 to set the desired temperature (see below for using Apex3). If the unit is switched off it needs to be started. Press the main single button on the front of the controller. Follow the on screen prompts to switch on the unit. Use the <Run Experiment> interface of Apex3 to set the desired temperature (see below for using Apex3). Starting the System Start BIS (Bruker Instrumentation System) if it is not already running. This will also start the measurement server if it is not active already. 5

6 Figure 3, The BIS window Unloading the Previous Sample Check if a data collection is still running (e.g., look at the BIS interface, check for Frames left or Time left ). If a data collection is active but can be aborted, then stop this data collection first. Navigate to the active copy of Apex3 and click the STOP symbol in the Apex3 interface. Minimize any sessions of Apex3 that might be open. Start a new session of Apex3. No login is required. Connect to BIS (<Instrument>, then <Connect>). Via the <Sample> dropdown menu, either <Open > an existing sample (continuation of old data collection) or Start a <New > Sample Fill in the project name. Avoid using special characters (including dashes and periods) or overly long project names. Under <Set Up>, go to <Center Crystal> Click on <Center> in the lower right of the interface. The angles will drive to a position convenient for mounting a magnetic snap-on mount. 6

7 Press the <door open> button on the front face of the diffractometer enclosure (to the right of the front door, below the light button). Doors need to be closed to drive any mechanical parts other than the phi (spindle) axis of the goniometer. Carefully dismount the magnetic snap-on mount from the previous experiment. Tilt it backwards until the magnet comes loose, then take it off. Figure 4, Goniometer head with magnetic snap on mount and adjustment screw Selecting and Mounting of Sample Crystals can be mounted on glass rods, inside glass capillaries, on Nylon loops, or using mesh mounts. The Purdue X-ray lab uses Mitegen micromesh mounts for most samples and data collections. Two types of micromesh mounts for small (<0.1 mm) and medium and large ( mm) crystals are stocked in the lab (Fig 5). Crystals collected at low temperature are mounted on the micromesh with the help of a trace of mineral oil, Fomblin (fluorinated mineral oil) or polybutene oil (very viscous oil to protect extremely oxygen sensitive samples) and flash 7

8 frozen in the cold stream. Crystals collected at room or elevated temperature might need to be fixed to the mesh with a trace of glue to avoid crystal movement during measurement. Figure 5, Tips of Mitegen Micromesh mounts as seen through a microscope (Left: small tipped mount with 15 µm openings for crystals 0.1 mm and smaller. Right: large mesh mount with 25 µm openings. Mesh is 0.4 mm across). For typical crystal selection and mounting, place a drop of mineral or Fomblin oil on a glass slide under a microscope. Place several representative crystals in the oil. Inspect the crystals and select a suitable candidate. Crystals on the Quest instrument with copper microsource should ideally be not larger than mm in any direction (the size of the X-ray beam is 0.1 mm). The minimum size depends on the diffraction intensity of the crystal. Larger crystals, especially if they are heavily absorbing, are usually more suitable for the Quest instrument with molybdenum sealed tube X-ray source (the size of its X-ray beam is 0.6 mm). Data collections are much faster using molybdenum radiation as long as the achievable diffraction is sufficiently intense. The higher intensity microscource and the intrinsically better diffraction with copper radiation is ideally suited for small weakly diffracting crystals that are unsuitable for molybdenum radiation. Only those samples and organic samples for which absolute configuration is required should be run on the copper wavelength instrument. For highly absorbing samples limit the size of the crystal (make sure the minimum transmission will be above 10-20%). If no single specimen with suitable dimensions can be found, use a sharp razor blade to cut off a single piece from a larger crystal or cluster. Remove all smaller pieces and dirt from the crystal (the micromesh mounts can be used to brush off loose pieces from the crystal). 8

9 Figure 6, Crystals with well-developed faces on a microscope slide Scoop up a crystal with the mircomesh mount and place it on the inside of the concave face of the mount in the center of the mesh. Remove as much excess oil as possible. (If a crystal needs to be glued to the mount make sure that both crystal and mount are dry without any traces of oil. Dip the mesh carefully in a small amount of glue and pick up the crystal without touching the glass slide). Crystal Mounting and Centering At this point you should have a session of Apex3 running. Press the <door open> button on the front face of the diffractometer enclosure (to the right of the front door, below the light button). Carefully mount the magnetic snap on mount onto the goniometer. Check the position of the mount by eye (is it aligned with the collimator and the beam stop?) In Apex3, click <Center Crystal> in the <Set Up> submenue. Click the blue-green play button at the top of the window to open the live video stream feed. Is the crystal visible on the video screen? If you cannot see the crystal, its position is further off than usual and needs to be adjusted by eye on the goniometer head before the live video feed can be used. 9

10 Figure 7, Crystal Centering Window and live video feed with mounted sample and green measuring line. There are three pins on the goniometer head: for up-down, right-left, and forward-backward. Use the bold end of the adjustment screw driver to turn the pins to position the crystal in the center of the cross hair of the video feed. When the crystal is centered in the cross hair, click the large <Spin Phi 90> button (in the <Center Crystal> window of Apex3) to spin the crystal around the phi axis by 90. Repeat the centering procedure using the adjustment screws as before. Repeat the process (<Spin Phi 90> followed by centering of the crystal in the cross hair of the video feed) until the crystal stays positioned in the center of the cross hair in all positions. If the cross hair seems to be slightly off try to center the crystal so that it rotates in place. For oddly shaped crystals, try to center the mass center of the crystal. For long needles, try to align them along the axis of the mounting pin. Measure the dimensions of the crystal: Switch on the measuring tool (double sided arrow icon on top of <Center Crystal> window, Fig. 7). Click on one side of the crystal, hold the left mouse key and drag to the other side. The length of the green line dragged is given in the 10

11 bar below the video screen ( vector length, value in µm). Write down the value and repeat for the other two directions using <Spin Phi 90>. Crystal and Compound Description Under <Setup>, go to <Describe> Fill in all values that apply. In the compound line, it is best to give the full name of the person that provided the sample. Crystal dimensions need to be entered in mm (video screen values are in µm!). Figure 8, Crystal and compound description 11

12 Unit Cell Determination Under <Evaluate>, click on <Determine Unit Cell>. Two procedures are available, <Automatic Mode> or <Manual Mode>. Figure 9, Starting Window of the Unit Cell determination We encourage to use the manual procedure. To do so click <Collect Data>, this will open the Unit Cell Data Collection Window. Choose an exposure time (default is 10 seconds) and click <Collect> frames will be collected which will be displayed in the frame window. The first sweep will be at low angle, with the beam stop visible at the center (Figure 10). The second and third run (Figure 11) will be at high angle. Make sure that your specimen diffracts to sufficiently high angle. Use the resolution arc tool (Figure 11) to check the resolution limit. Only samples that show at least weak diffraction above ca. 0.9 Å (with ten second exposures) are likely to meet the IUCr requirements for publication. 12

13 Figure 10, Unit Cell window with the first (low angle) sweep running. Figure 11, Unit Cell window with one of the high angle sweeps running and resolution arc tool. 13

14 When finished proceed to <Harvest Spots>, select an I/sigma(I) cutoff value for the diffraction spots to be used (see the frame to the left, spots that will be used are green circled). To obtain a reasonable suggestion for the exposure time use the default value of Click on <Harvest>. Figure 12, Unit Cell Harvesting Window Proceed to <Index>, use all default values and click on <Index>. 14

15 Figure 13, Unit Cell Indexing Result Window Check the unit cells obtained. The positions of the diffraction spots should agree with the predicted positions (white or blue circles). The hkl histogram (on the right) should have values around 90% or higher in the 0.1 line. If neither of the two is the case a nonmerohedral twin might be present, or the crystal is not single. If this is the case, export the data as a p4p file using <Sample>, then <Export> and use the program Cell Now to determine the cell (see the Purdue Twinning Manual for details). If one of the two unit cells looks reasonable, select it and proceed to <Refine>. Use all defaults and click on <Refine>, then <Accept>. 15

16 Figure 14, Unit Cell Refinement Window Proceed to <Bravais>. A list of possible Bravais lattices will be displayed with the software's choice highlighted in green. 16

17 Figure 15, Bravais Result Window The correct choice should have an FOM (figure of merit) value significantly higher than the others. If several solutions have similarly high FOM values, the one with the highest symmetry is likely to be correct. If you are not sure (e.g. when a high symmetry solution has a high FOM value, but significantly lower than a lower symmetry solution) select the lower symmetry solution or your data set as set up in <Data Collection Strategy> might be incomplete. Proceed to <Refine>, click <Refine>, then <Accept>. Please note the resolution prediction at the bottom of the window This finishes the Unit Cell Determination. 17

18 Data Collection Click < Calculate Strategy> under <Collect>. Under Step 1., change the resolution to d = 0.82 Å or smaller. The worst acceptable value, by IUCr standard, is 0.83 Å. The best resolution that can be set on this instrument is 0.78 Å. Collecting to 0.78 rather than 0.82 significantly increases the data collection time and is only recommended for samples that do indeed intensely diffract all the way to the detector edge. Click <Apply> if you made any changes. Figure 16, Strategy Starting Window Under Step 2., click <Determine Strategy>. In the pop-up window, it is usually save to use the default values. Click <OK>, 18

19 Figure 17, Determine Strategy popup window Under Step 3., click <Select Scan Parameters>. A popup window will emerge with suggested values for frame angle and frame time. Figure 18, Strategy Window with Scan Parameters sub-window The values can usually be accepted as they are. 19

20 Under <Collect> click on <Experiment> Click <Append Strategy> to add the runs from the Strategy calculation. On this instrument, usually no Fast Scans are required. Instead, edit the exposure times for the low angle runs (runs with values < 30 degrees in the 2Theta column) to one half or one quarter of the other exposure times. For medium angles (30 60 degrees, use 2/3 rd of the suggested exposure time. This avoids overexposure of the detector for the more intense low and medium angle diffraction spots. The long exposure as estimated in strategy should be retained for high angle runs (> 60 degrees in 2Theta). To preliminarily solve your structure usually each one low angle and one high or medium angle run are needed on this instrument. If the first low angle run is scheduled by strategy rather late into the measurement then it might make sense to move it up in the order of runs. Click on the run you would like to move so that it is highlighted in dark blue. Move the mouse cursor over the number to the left of the line and right click. Use <cut> to remove the line. Highlight the first line, right mouse click on the number to the left again and use <paste> to move the run into first place. 20

21 Figure 19, Experiment setup window during editing of low angle exposure time Click <Execute> to start the data collection. The instrument will now collect your dataset. Open <Show Status> from the Instrument drop down menue. This will open the <Instrument Status> window where you can check variables and completion time. You can also check many of these items via the BIS interface. 21

22 Figure 20, Data Collection Window with active data collection and BIS Window in foreground Users are encouraged to check the progress of their data collection and to test-solve and refine the data while the data collection is still running to check that the data quality is good enough to proceed, that the structure is indeed of interest, that the unit cell is correct, and if data might be already complete. It is highly advisable to not stop a data collection until you are sure the data are complete enough for a meaningful absorption correction and to pass checkcif (i.e. data should be integrated, scaled, solved, refined, the refinement quality checked, the completeness should be checked in the cif file, and preliminarily checked using checkcif prior to stopping the data collection!!). Once a crystal is taken off the instrument a data collection cannot be resumed! 22

23 Data Integration When the data collection is complete (or when enough data are collected for an initial structure solution and refinement attempt) open, <Reduce Data>, then <Integrate Images> Change the resolution limit to 0.83 Å or lower Click on <Find Runs>, check the runs you would like to integrate (leave out the Matrix runs). Figure 21, Integration Start Window with <Select Runs> Window open Click on <Start Integration>. This will start the integration. Wait for it to finish, this might take several minutes. 23

24 Figure 22, Integration Window with active Integration running Spot Shape Correlation should show values around or above 80% and be mostly even. Values of 40% or lower are statistically meaningless. Pixel Errors should be on average between ±0.2. Average I/sigma(I) intensities should be above at least 3 (good datasets have values larger than 20). Shape profiles should be round to slightly ellipsoid. Double spots, large pixel errors and erratic uneven lines are possible signs for twinning. If you suspect a crystal to be split or twinned, consult the Purdue Twinning Manual. If you encounter problems with data completeness: Open the _0m._ls file and inspect the table at the beginning of this file. Inspect especially the values for Spots exceeding frame queue size and Spots left in Write-Behind Cache. If these are unusually large (a substantial fraction the total data), then click <Integration Options> and change the values for Profile XYX Half-Widths from 4 x 4 x 4 to 9 x 9 x 9. Repeat the integration. If you used very narrow frame widths you might also want to increase the number for Active Image Queue Half-Widths (e.g. from 7 to 15, maximum is 32). 24

25 Scaling and Absorption Correction Under <Reduce Data> Click on <Scale> Figure 23, Absorption Correction and Error Model (<Scale>) Starting Window The merged batch raw file from the integration is loaded automatically. If you did not integrate all runs at once you need to load individual raw files. Click the yellow browse symbol, click on any raw file you want to load, then click <OK>. Uncheck raw files you would like to not use. Check the Laue Group and Point Group. Keep Multi-scan as the absorption correction method. It is sufficient for > 99% of all samples. For heavy absorbers, adjust the number of Mu*r (mu is the absorption coefficient in mm -1, r the average radius of the crystal in mm). Proceed by clicking <Next>. 25

26 Figure 24, Absorption Correction window (second Scale Window) before refinement The refinement window will open (Fig. 24). An average redundancy of around 5.0 is usually sufficient for weak to medium absorbing samples. For oddly shaped and heavily absorbing samples a higher redundancy is recommended. Click <Refine>, wait for the refinement to finish and inspect the results (Fig. 25). 26

27 Figure 25, Absorption Correction window (second Scale Window) after refinement Click <Next>. Check the number of rejected reflections. If this number is extremely high, in conjunction with higher than expected data R values, then there is a possibility that the selected Laue group is too high in symmetry ( metric pseudosymmetry ). The number of rejected reflections can be reduced (assuming that the Laue group is correct!) by increasing the value for < I-<I> /su ratio for rejection> (default is 4.0). If you do so click <Error Model> to update the outlier rejection. 27

28 Figure 26, Error Model Window (third Scale Window) Click <Finish>, then <Exit> to finalize scaling and absorption correction. You have now finished the data collection and absorption correction. To allow other users to use the instrument you should move your data to a different computer at this point. Transferring the Data Move the data via a USB flash drive or online to the data workup computer or your personal computer. For the refinement, you will need the *.hkl and the *.p4p files (located in the work folder of your project) e.g. 04mz02a_0m.hkl 04mz02a_0m.p4p For publication purposes you will also need the *.abs and the *._ls files: e.g. 04mz02am.abs or 04mz02a_0m.abs (A copy of what you did in SCALE, SADABS, or TWINABS, contains the ratio of Tmin/Tmax) 28

29 04mz02am_0m._ls (A copy of the last lines of the integration procedure, contains parameters of unit cell refinement (2THETA min, 2THETA max and the number of reflections used), crystal colour and shape, crystal dimensions) If you used Cell Now to obtain a unit cell, you will also need to copy the *._cn file Setting up Data for XPREP On your computer save the files from above into a new file folder. Making a backup of your files is strongly recommended! (with older operating systems you need to uncheck the read only flag, by highlighting your files plus right mouse click, go into properties and uncheck.) Open the SHELXTL program. Select <PROJECT> and <New>. Find the appropriate file and open it. Give it a project name e.g. 04mz02a_0m, then <open>. Figure 3, Shelxtl Program and Project Manager Window with New Project Window open 29

30 XPREP Select <XPREP> on the toolbar. In the next steps, the computer will make suggestions that can usually be accepted (i.e. for high quality data). Figure 4, XPREP Window, initial lattice centering selection 30

31 - Select the suggested lattice type. - Choose [H] to search for higher metric symmetry. - Choose offered choice [A] for the Laue group (e.g. monoclinic). Figure 5, XPREP Window, initial lattice type selection - Select [S] to determine or input space group. - Select [S] again to determine space group. - Select the suggested Laue group (e.g. [M] for monoclinic) 31

32 - Select the suggested lattice centering (e.g. [P] for primitive, [C] for C-centered, etc) Figure 30, XPREP Window, new lattice centering selection - Select the suggested option for the space group e.g. P2(1)/c or C222(1) (If several solutions are offered, take the one with the lowest CFOM value. If that does not work out later on, try the next best in the list). 32

33 Figure 31, XPREP Window, space group selection - Select [D] to read, modify or merge datasets - Select [S] to display the intensity statistics - Select [A] to merge all equivalent reflections (including Friedel opposites). This does not actually merge the reflections (which is not recommended to do), but only displays the data and statistics as if they would have been merged. 33

34 The intensity statistics will be displayed. Figure 32, XPREP Window, intensity statistics Have a look at the Completeness, Redundancy, Rmerge and Rsigma (Rmerge is the same as Rint in Shelxl and Sadabs). The completeness should be ideally close to 100 % down to a resolution of d = 0.82 Å. Rsigma and Rmerge should be ideally below 10 % down to d = 0.82 Å. Hit the Enter key - Select [E] to exit to the main menu, then [C] to define the unit cell contents. - If no unit cell contents are given type the most likely molecular formula (element symbols and numbers only). - If the given formula seems wrong, select [F] for new formula, and type the most likely formula. 34

35 Figure 33, XPREP Window, formula definition - If necessary, select [R] to change the radiation (Mo radn; λ = ; Cu radn; λ = ) - If necessary, select [Z] to change the number of (symmetrically independent) molecules Z per unit cell. - Select [E] to exit to the main menu. Note: If you had been using SADABS for the generation of your *.hkl file, no absorption correction has to be applied. If you still need to apply absorption correction, this can be done here by choosing [A] (not covered here). - Select [F] to setup the new hkl file. If you have changed the unit cell or its orientation in XPREP, the program will force you to choose a new name for the hkl file to avoid overwriting and losing the original data (change it back to your old filename after closing XPREP, or start a new project with the new hkl filename). - Select [Y] at the prompt to generate an.ins file. 35

36 Figure 34, XPREP Window, ins and hkl file setup - Select [Q] at the prompt 36

37 The Initial INS File The initial *.ins file written in XPREP, together with the *.hkl file, contains all the information to get started solving and refining your crystal structure. The *.ins file can be opened and edited in any text editor. Open it via <Edit> and <Edit.ins> in the SHELXTL menu. The initial ins file may look like this: Figure 6, typical initial INS file The TITL line gives the initial name of the *.ins file and the space group from XPREP. The cell line gives the wavelength used, then the unit cell parameters a, b, c, α, β, γ. The ZERR line gives the Z value (number of molecules in the unit cell as given in XPREP) and the estimated standard deviations for a, b, c, α, β, γ. LATT gives the lattice type (see SHELXTL manual). The SYMM line(s) give the symmetry operators of the space group that create symmetry dependent atoms (these lines define the actual space group, not the TITL line). The SFAC line gives the atom types and defines the atom structure factors to be used by the program. The UNIT line gives the number of atoms per unit cell as listed in the SFAC line. The TEMP line gives the temperature in degrees Celsius. If needed, change the value to the actual temperature used in your experiment. The TREF line defines the type of structure solution method to be used (see below). 37

38 The HKLF 4 line defines the type of refinement to be used (HKLF 4 is the standard single crystal refinement for a not non-merohedrally twinned dataset). The END lines defines the end of the file. Select <file> and then <save>. Solving Structures using XS and XM To perform the refinement of your structure, you will first need an initial guess (a starting point). Using the *.ins file provided by XPREP, the programs XS or XM will provide you with this initial guess (other programs and methods can be used, such as DirDiff, charge flipping, etc). In the program XS there are two methods to choose from: Direct methods (indicated by the line TREF in the INS file), and the Patterson method (indicated by the line PATT in the ins file). Direct methods can be used for all datasets with better than atomic resolution. Patterson methods can only be used for structures with at least one heavy atom (i.e. sodium or heavier). Direct methods usually provide a more complete initial guess (more atoms already found), but they tend to fail more often than Patterson methods and they have occasionally problems to place inversion centers correctly. XM is utilizing direct methods originally written for macromolecular data, but can also be used for small molecule data. It is more likely to work where XS/TREF fails, but requires more computing resources. - To use XS, select <XS> on the toolbar and the computer begins to process data. The software tries in this step to find an initial guess for the atom positions. This can only be successful if the atoms listed in the formula are correct. The software will select the solution with the smallest value for CFOM (combined figure of merit) and the initial atom positions are written to the *.res file. The res file cannot be opened while XS is running. 38

39 Figure 7, the XS window - For good quality data the correct solution is normally falling out when continuing with Shelxle, Olex or Crystals graphical refinement interfaces. If this is not the case, you can change the settings for XS in the ins file. Go to <Edit> on the toolbar and select <edit.ins> Either, change TREF to TREF 2000 (or even higher) to run more iterations, or, for compounds with at least one atom heavier than sodium, use PATT instead of TREF. Select <file> and then <save> and run XS again. If XS fails using both PATT and TREF methods, try using <XM>. In the *.ins file, replace the TREF line by the following four lines: FIND 58 PLOP 77 MIND NTRY

40 Save the ins file and run Click <XM> in the in the SHELXTL menu. The software will run 1000 tries (or the number specified in the NTRY line) and selects the solution with the highest correlation coefficient, CC. The initial atom positions are written to the res file. This will take substantially longer than XS (a fast computer may be required), but solutions can be analyzed while XM is still running. If a valid solution is found, XM can be aborted. XM does not write all required lines to the *.res file. Before running the first refinement cycle, add the following lines: FMAP 2 L.S. 4 PLAN 20 Structure Refinement Open the result of the structure solution in a graphical interface such as Shexle, Olex or Crystals. See separate manuals for details of the refinement procedure, and for writing and validating the crystallographic information file. 40